Revert to simpler single output text streamer

Try to fix by removing skip special token

Fix bugs

Fix end on "User:"

Break on "User:"

fix indentation

Try to fix

Fix richtextstreamer for multiple tokens input

Free to choose a model

Implement reasoning blocks and fix eos token showing

Implement richtextiteratorstreamer

Put everything on the cuda device!

Fall back on avoiding '\nUser'

Remove custom stoppingcriteria and trust generate

Set pad id

set pad and eos tokens

debugging

Do not skip special tokens

Fix stopping criteria

Stop at eos token

Add threading import

Use background threading for generation

Remove one chat_with_model function

Do sample

Add break after end of sequence token

Try to fix chatbot streaming

Change chatbot

Make it into a chatbot

Fix loading on startup

Fix streaming

Fix for streaming output

Generate with enter

Fix model to cuda

Put model on GPU

Load models using dropdown

Fix loading of models (only once per model)

Show loading bars

Install accelerate

Force install transformers

Add requirements.txt

first commit

initial commit